How much do you smoke per day (cigarettes/cigars etc.)?
Available answers were: None, 1 – 10, 11 – 20, or >20

Statistical Methods

Patient demographics were summarised by CRS diagnosis status using mean and standard deviation for continuous variables and the number and percentage for categorical variables. The mean SNOT-22 scores were compared between active smokers and non smokers in each CRS diagnosis group and the control group using a two sample t-test. A linear regression model was used to test if the difference in mean SNOT-22 score between active smokers and non smokers depended on the CRS subgroups using a test of interaction between CRS subgroup and smoking. No difference was detected and hence the three groups were combined into a single analysis. A linear regression model was used to adjust for potential confounding due to age, gender and a diagnosis of asthma. All analyses were conducted using Stata MP 16.0.

Results

Study Participants

A total of 1535 questionnaires were returned with 1470 considered eligible for inclusion after removal of duplicates and questionnaires with missing data (see figure 1). The overall response rate of those identified to take part in the study was 66% of those distributed. This analysis is therefore based on the 1450 participants who completed the smoking part of the questionnaire.

Descriptive Data

For the purpose of this analysis, patients with AFRS and CRSwNPs are categorised together as a single group (CRSwNPs+). As such, there were 219 controls, 546 participants with CRSsNPs and 685 participants with CRSwNPs+. The patient demographics are outlined in Table 1. With similar mean age between groups and greater female preponderance in the control group and male in the CRSwNp group.

Primary Outcome Data and Main Results

The majority of active smokers in both control and CRS groups reported smoking less than 10 tobacco products a day (63% and 61% respectively). Comparing disease groups there appears a greater number of active smokers amongst controls (15%), which itself was below the 2007-2009 UK national average of 21%.20 Table 2 outlines the differences in the rates of active smokers between the three groups (p=0.039, Chi-squared test) and highlights the lower rate of smokers in CRSwNPs+ participants (9.9%) and CRSsNPs patients (13.9%) respectively (p=0.03, Chi-squared test).
Mean SNOT22 scores were notably higher in the smoking cohorts for all three phenotypes. On calculating the mean difference in SNOT-22 score between active smokers and non-smokers we found a significant difference for both CRS phenotypes (p<0.001 on Analysis of Variance (ANOVA), see Table 3). In both CRSsNPs and CRSwNPs+ groups, active smokers had significantly worse SNOT-22 scores than non-smokers by a mean magnitude of 10 points. This remained significant after adjusting for age, sex and asthma (Tables 3 and 4).
Categorising CRES participants by smoking preference demonstrated a higher percentage of surgical procedures within the non-smoking cohort (Table 5), however there was no statistical difference between smoking and reporting multiple (1 or more) surgical procedures (p=0.098).